Psychosis is a construct that has classically been treated as the ?hallmark feature? of schizophrenia, but is nonetheless present in other disorders including schizoaffective disorder and bipolar disorder. Persons with psychosis experience profound and disabling dysfunction in multiple domains of function, including motor system, cognitive control, and social function. This exacts a tremendous toll on afflicted persons, their communities, and society at large. Due to definitional issues, any two persons with psychosis may have vastly different psychotic features. This leads to significant within-diagnosis heterogeneity, which may contribute to the lack of disease specific diagnostic findings. In addition, there is significant comorbidity across diagnoses, which potentially results in under-sampling of the full range of disease expression when only one diagnostic group is included in a given study. To address these issues, the NIMH developed the Research Domain Criteria (RDoC) project to link behavioral endpoints with their underlying neural mechanisms. In line with the RDoC goals, the proposed project focuses upon determining dimensional neural predictors of three recognized domains of dysfunction in psychosis using a transdiagnostic subject population comprised of persons with schizophrenia, schizoaffective disorder, and bipolar disorder. To accomplish this, multiset canonical correlation analysis + joint independent components analysis (mCCA+jICA) will be employed to analyze a multimodal neuroimaging dataset comprised of gray-matter, white matter, and functional connectivity measures using a meta-analytic approach to selecting areas empirically identified as related to three psychosis domains of interest. Methodologies will be prototyped using data from the Human Connectome Project (HCP) and then extended to a dataset comprised of persons with psychosis. For each specific aim, analyses will: 1) select meta-analytically determined areas; 2) extract gray-matter, white-matter, and functional connectivity measures; 3) perform mCCA+jICA to identify the significant orthogonal sources of variance; and, 4) use a regression approach to link domain-specific behavioral and self-report measures to the sources identified by mCCA+jICA. In Aim 1, motor system dysfunction will be examined and linked to mCCA+jICA results using two measures of the motor system. In Aim 2, cognitive control dysfunction will be examined and linked to mCCA+jICA results using a composite measure of cognitive control and then followed-up with four individual measures. And, in Aim 3, social dysfunction will be examined and linked to mCCA+jICA results using a composite measure of social function and then followed-up with four individual measures. The proposed research may lead to findings that better support dimensional models of dysfunction within domains of psychosis and better modeling of psychopathology. Further, it may allow future research efforts to study more homogenous subject populations enabling more targeted diagnosis and treatment of specific dysfunction in psychosis.